Billed as "Twitpic for Audio", the Twaud.io service (Twitter Audio) allowed a short-form URL to post audio snippets on Twitter. Launched in May of 2009 by Massive Robot, Twaud.io was one of a number of third-party services bringing rich content access to Twitter streams. With a limit of 10 megabytes, no limit on content or approach, and an easy to use API, Twaud.io seemed poised for some level of success.
In 2011, Twaud.io announced it was shutting down, and gracefully cut off uploading of new content while providing easy to use download links for users and months of warning. This dataset represents roughly 21,000 mp3s from 4,300 users and related metadata from publicly-available postings on twaud.io across its 3 years of operation.
Functioning as a social experiment and as a repository, this archive of twaud.io contains speeches, songs, interviews, random noise, and additional varieties of twitter-inspired audio collected from all over the world and all over time. It is encouraged as a sociological study and as an insight into the use of social media.
A .txt file of the contents of this 60 gigabyte archive is included for searching, although the filenames are non-informative.